Nephrol Dial Transplant (2013) 0: 1–10doi: 10.1093/ndt/gft399

Original Article

Clinical and immunological implications of increase in CD208+

dendritic cells in tonsils of patients with immunoglobulin Anephropathy*

Hanako Takechi1,

Takashi Oda1,

Osamu Hotta2,

Kojiro Yamamoto1,

Naoki Oshima1,

Takeshi Matsunobu3,

Akihiro Shiotani3,

Hiroshi Nagura4,

Hideyuki Shimazaki5,

Seiichi Tamai5,

Yutaka Sakurai6

and Hiroo Kumagai1

1Department of Nephrology and Endocrinology, National Defense

Medical College, Tokorozawa, Saitama, Japan,2Hotta Osamu Clinic, Sendai, Miyagi, Japan,3Department of Otolaryngology, National Defense Medical College,

Tokorozawa, Saitama, Japan,4Sendai Red Cross Hospital, Sendai, Miyagi, Japan,5Department of Laboratory Medicine, National Defense Medical

College, Tokorozawa, Saitama, Japan and6Department of Preventive Medicine and Public Health, National

Defense Medical College, Tokorozawa, Saitama, Japan

Correspondence and offprint requests to:Hanako Takechi;

E-mail: DRI2437@ndmc.ac.jp*Part of this work was presented at the annual

meeting of the American Society of Nephrology in 2012.

Keywords: CD208, dendritic cells, glomerular crescent, IgAnephropathy, tonsillectomy

ABSTRACT

Background. The therapeutic effect of tonsillectomy for im-munoglobulin A nephropathy (IgAN) has been widely recog-nized, but the mechanism by which tonsillar immunity leadsto glomerulonephritis has been unclear. We investigated sub-types and localization of dendritic cells (DCs) in tonsils andlooked for relationships between the tonsillar DCs and theclinical features and renal histological changes of patients withIgAN.

Methods. We examined tonsils from 33 IgAN patients, using ascontrol tonsillar specimens from subjects without glomerulone-phritis. Five distinct markers of DCs (CD303, CD1c, CD209,CD208 and CD1a) were analyzed by immunohistochemistryand flow cytometry. The mRNA levels of these DC markers wereevaluated using real-time polymerase chain reaction. The clinicaldata and histological results obtained evaluating renal biopsytissues were statistically compared with immunological data.Results. Of the five subtypes of DCs, CD208+ DCs were sig-nificantly increased in the tonsils of IgAN patients comparedwith that of controls. Furthermore, the number of CD208+

© The Author 2013. Published by Oxford University Press on behalf of ERA-EDTA. This is an Open Access article distributed under the terms of the CreativeCommons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/), which permits noncommercial re-use, distribution, andreproduction in any medium, provided the original work is properly cited. Forcommercial re-use, please contact journals.permissions@oup.com

1

NDT Advance Access published September 29, 2013 by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

by guest on N

ovember 3, 2013

http://ndt.oxfordjournals.org/D

ownloaded from

DCs in the tonsils was positively and linearly correlated withthe proportion of crescentic glomeruli in renal biopsy tissuesand with the urinary protein level. Only few CD208+ cells,however, were found in the kidney biopsy specimens of IgANpatients.Conclusions. These observations suggest that increasedCD208+ DCs in tonsils may play a directive role in the patho-genesis of IgAN. The present results support the therapeuticsignificance of tonsillectomy for IgAN patients.

INTRODUCTION

Immunoglobulin A nephropathy (IgAN) is the most commonprimary glomerulonephritis in the world [1, 2] and has beenfound to have a worse renal prognosis than had originallybeen expected [3]. There is, therefore, a need for the effectivetreatment, which results in remission of the disease in its earlystage.

Hotta (an author of this manuscript) et al. [4] first advocatedthe effects of treating IgAN patients with a combination of ton-sillectomy and steroid pulse therapy not only on reducing he-maturia/proteinuria, but also on long-term remission regardingrenal function. A recent prospective randomized controlledstudy also showed 2-year favorable effects of the combination oftonsillectomy and steroid pulse therapy on proteinuria andrenal function [5]. Moreover, efficacy of tonsillectomy for therecurrence of IgAN after kidney transplantation has been re-ported [6, 7]. Although these clinical observations stronglysuggest a relationship between the tonsillar immune system andthe pathogenesis of IgAN, the mechanism by which tonsillarimmunity leads to glomerulonephritis is unclear.

Dendritic cells (DCs) are the most potent antigen-presentingcells, and their participation in the pathogenesis of IgAN hasbeen suggested in some reports. Deterioration of renal lesionswas observed after mucosal activation of DCs in the animalmodels of IgAN [8], and the expression of B-cell activationfactor on DCs has been reported to have been significantlyhigher in IgAN patients than that in controls [9]. These reportsindicate that DCs play key roles in the pathogenesis of IgAN.

Different subtypes of human DCs have been classified ac-cording to their phenotypic characteristics. All DCs derivefrom hematopoietic stem cells, which differentiate into varioussubtypes of DCs along two main branches: a series of lym-phoid precursors and a series of myeloid precursors. The lym-phoid precursors give rise to the plasmacytoid DCs (pDCs)expressing blood DC antigen 2 (BDCA2), alias CD303 [10],whereas the myeloid precursors give rise to conventional DCsexpressing blood DC antigen 1 (BDCA1), alias CD1c, mono-cyte-derived DCs (moDCs) expressing DC-specific intercellu-lar adhesion molecule 3 grabbing, non-integrin (DC-SIGN),alias CD209, and Langerhans cells expressing CD1a (mainlylocalized in skin) [11, 12]. Interdigitating DCs (iDCs), expres-sing DC lysosome-associated membrane glycoprotein (DC-LAMP), alias CD208, mainly localize in the interfollicular areaof lymphoid tissues. CD208+ DCs are thought to be derivedmainly from moDCs [13].

So far, it has not been shown which subtypes of DCs takepart in the pathogenesis of IgAN. We therefore investigated thesubtypes and their localization in the tonsils of patients withIgAN and looked for correlations between the types of tonsillarDCs and the patients’ clinical and renal histological parameters.

MATERIALS AND METHODS

Patients and controlsThe study protocols were approved by the Ethical Committeeof National Defense Medical College.

Thirty-three of the biopsy-proven IgAN patients who hadundergone tonsillectomy from January 2008 to September2012 in the National Defense Medical College Hospital par-ticipated in this study. They were selected by excludingpatients tonsillectomized >1 year after renal biopsy, patientssuffering from diabetes, neoplasm, inflammatory diseases orother systemic diseases and patients who had received corti-costeroids and/or other immunosuppressive agents. Also par-ticipating were nine control patients who had no history ofrenal, liver, systemic diseases and active infectious diseases. Allcontrol specimens were taken from the chronic tonsillitispatients without glomerulonephritis (without proteinuria orrenal dysfunction). Written informed consent was obtainedfrom each patient in accordance with the principles of the De-claration of Helsinki. Characteristics of both the IgAN andcontrol patients are listed in Table 1.

Histological evaluationSections (3 μm thick) of formalin-fixed, paraffin-embedded

tonsillar tissues were stained with hematoxylin and eosin(HE). For the immunostaining of CD209 and CD208, two-step indirect immunoperoxidase staining was performed usingthe secondary antibody conjugated with the horseradish per-oxidase-labeled polymer (Envision; DAKO Corp. Carpinteria,CA, USA) and developed with 3,30-diaminobenzidine reagent(Thermo Fisher Scientific, Inc., Waltham, MA, USA). We per-formed immunostaining of CD208 with the use of antigen re-trieval process of proteinase K (DAKO Corp.) digestion for 10min and subsequent heating at 95°C in citrate buffer (pH 6.0)for 20 min. The counterstaining was performed with Mayer’shematoxylin (Wako, Osaka, Japan).

Direct immunofluorescence (IF) staining for CD303, CD1cor CD1a was performed on acetone-fixed, frozen tonsillar

Table 1. Characteristics of IgAN patients andcontrol patients

Controls IgA nephropathy

Number of patients 9 33

Sex (male/female) 1/8 9/24

Age 41.7 ± 20.4 33.9 ± 17.7

Serum Cr (mg/dL) 0.61 ± 0.18 0.73 ± 0.22

Urinary protein (g/gCr) <0.1 0.70 ± 0.98

ORIG

INALARTIC

LE

H. Takechi et al.

2

sections (5 μm thick) using fluorescein isothiocyanate (FITC)-conjugated primary antibodies. Triple IF staining for CD208,CD3 and HLA-DR and for CD20, CD3 and IgA were also per-formed on frozen tonsillar sections of several patients withIgAN. In the process of triple staining, staining for IgA wasperformed with a direct IF staining method, and staining forCD208, CD3, HLA-DR and CD20 were performed with an in-direct IF staining method using secondary antibodies; AlexaFluor 350-labeled goat antirabbit immunoglobulin G (IgG;Molecular Probes, Eugene, OR, USA), Alexa Fluor 594-labeleddonkey antimouse IgG (Molecular Probes) or Alexa Fluor488-labeled donkey antirat IgG (Molecular Probes). IndirectIF staining for CD208 was performed on frozen sections(3 μm thick) of renal tissues of patients with IgAN (n = 17)using secondary antibody; Alexa Fluor 488-labeled goat anti-mouse IgG (Molecular Probes). Primary antibodies used inthis study are listed in Table 2.

To quantify the area of germinal centers of tonsils, we as-sessed HE-stained sections under light microscopy. Photo-graphs of a random field were obtained using a digital camera(magnification; ×10), and germinal center area per total tonsil-lar area was measured by using the image analysis software(Lumina Vision Ver.2.04; Mitani Corp., Fukui, Japan). For thequantification of the number of each subtype of DCs, positivecells in the photographs of four consecutive fields (magnifi-cation; ×200) were counted and expressed as positive cellnumbers per square millimeter. Photographs of triple IF stain-ing were taken with a confocal laser scanning microscopy(LSM510; Carl Zeiss Co., Oberkochen, Germany).

Flow cytometryIsolated cells of tonsils from four IgAN patients and three

controls were analyzed by using flow cytometry as we reported

[14]. The fresh tonsillar tissues were minced into small pieces.After degrading them by adding 0.5 mg/mL collagenase(Wako), tissue debris was removed using mesh and they wereincubated with various antibodies listed in Table 2. Doublestaining was performed with antibodies for HLA-DR andCD303, CD1c or CD1a. Triple staining was also performedwith antibodies for Human Lineage Marker/HLA-DR Anti-body Mix and either CD209 or CD208. The lineage marker,which includes antibodies for CD3, CD14, CD16, CD19,CD20 and CD56, was used to exclude cell populations positivefor these markers from DCs in the analysis of CD208 andCD209, because there have been reports that a low level ofDC-LAMP (CD208) mRNA expression was observed uponactivation in other antigen-presenting cells such as monocytes,macrophages and B-cells [13], and that CD209 was also ex-pressed on CD14+ cells [15]. The hemolysis procedure wasdone using a Whole Blood Lysing Reagent Kit (BeckmanCoulter, Brea, CA, USA). Intracellular staining procedures forthe CD208 staining were performed using a BD Cytofix/Cyto-perm Fixation/Permeabilization Kit (Becton Dickinson andCompany, Franklin Lakes, NJ, USA) after staining cellularsurface markers according to the manufacturer’s instructions.Stained cells were analyzed by using the EPICS XL ADCSystem (Beckman Coulter). Data were analyzed using anEXPO32ADC software (Beckman Coulter). Antibodies usedfor flow cytometry are also listed in Table 2.

Real-time reverse transcription–polymerasechain reactionExtraction of total RNA, reverse transcription into compli-

mentary DNA (cDNA) and real-time polymerase chain reac-tion (PCR) using the 7900HT Fast Real-Time PCR System(Applied Biosystems, Foster City, CA, USA) were performed

Table 2. Antibodies used in this study

Antibody type Clone Labeling Supplier

CD303 AC144 FITC Miltenyi Biotec, Bergisch Gladbach,Germany

CD1c AD5–8E7 FITC Miltenyi Biotec

CD209 120507 Unlabeled, PE R&D Systems, Minneapolis, MN, USA

CD208 104.G4 Unlabeled, PE Beckman Coulter, Brea, CA, USA

CD1a HI149 FITC Biolegend, San Diego, CA, USA

Human Lineage Marker/HLA-DR Antibody Mix (hLMAX)

– Lineage; FITC, HLA-DR; PerCP-Cy5.5

Imgenex, San Diego, CA, USA

HLA-DR MEM-12YE2/36HLK

PEUnlabeled

Exbio, Prague, Czech Republic; GeneTex,Irvine, CA, USA

CD3 Polyclonal Unlabeled DAKO Corp., Carpinteria, CA, USA

CD20 L26 Unlabeled DAKO Corp.

IgA 106G FITC MBL, Nagoya, Japan

FITC, fluorescein isothiocyanate; PE, phycoerythin; PerCP-Cy5.5, peridinin chlorophyll protein complex with cyanin-5.5.

ORIG

INALARTIC

LE

I n c r e a s e i n C D 2 0 8 + D C s i n t o n s i l s o f I g A n e p h r o p a t h y

3

as previously described [16]. Total RNA from the tonsils of 23IgAN patients and 8 controls was analyzed. We used TaqManGene Expression Assays with primer/probe sets for humanCLEC4C (CD303), CD1c, CD209, LAMP3 (CD208), CD1aand glyceraldehyde-3-phosphate dehydrogenase (GAPDH;Applied Biosystems). The relative amount of mRNA was cal-culated using comparative Ct (▵▵Ct) methods. All specificamplification products were normalized against GAPDHmRNA, which for internal control was amplified in the samereaction.

Clinical and renal histological dataTwo pathologists evaluated the renal biopsy specimens in-

dependently. The diagnosis of IgAN was based on the demon-stration by direct IF staining of mesangial immune depositsthat stain dominantly for IgA, on mesangial proliferationevident by light microscopy and on the presence of mesangialelectron-dense deposits evident by electron microscopy. Cor-relations between the number of DCs in the tonsils and clinicaldata such as levels of proteinuria, levels of hematuria, serumcreatinine and serum IgA at tonsillectomy were assessed.Renal histological data such as proportions (to total glomeruli)of crescentic glomeruli, sclerotic glomeruli, glomeruli withadhesion and glomeruli with mesangial proliferation were alsoassessed in each renal biopsy, and their relations with thelevels of tonsillar DCs were evaluated. Renal biopsy evaluationand tonsillar tissue evaluation were done independently and inblind to each other.

StatisticsAll data were analyzed by using the JMP10 statistics

software (SAS Institute, Inc., Cary, NC, USA). Statistical ana-lyses were performed using Student’s t-test to compare themeans of two groups and Pearson’s correlation coefficient toshow the correlations between two variables. A P-value of<0.05 was considered statistically significant. Data are shownas mean ± SE.

RESULTS

Histological characteristics of tonsillar tissues in IgANLight microscopic observation of HE-stained tonsillartissues revealed that the interfollicular areas were characteristi-cally expanded, while germinal centers were reduced in tonsilsof IgAN patients compared with that of controls (Figure 1Aand B). This finding was quantified in HE-stained tonsillartissues by using an image analysis software to calculate theproportion of germinal center area to total tonsillar area. Theproportion of germinal center area to total tonsillar area wassignificantly lower in IgAN patients than that in controls(Figure 1C).

Immunohistochemical stainingsThe distribution of subtypes of DCs in tonsils did not basi-

cally differ between the IgAN patients and controls. In both,CD303+ cells were found sparsely in the under-epithelial andinterfollicular areas (Figure 2A). CD1c+, CD209+ and CD208+

cells were mainly found in the interfollicular area (Figure 2B–D). CD208+ cells were clustered in the interfollicular area(Figure 2D). CD1a+ cells were localized in the under-epithelialarea (Figure 2E).

As shown in Figure 3D, the number of CD208+ cells in thetonsils of IgAN patients was significantly greater than that inthe tonsils of controls. The levels of tonsillar CD303+, CD1c+,CD209+ or CD1a+ cells, however, did not differ between IgANpatients and controls (Figure 3A–C and E).

Triple staining for CD208, CD3 and HLA-DR revealed thatCD208+ cells often existed in touch with CD3+ T-cells. Fur-thermore, around CD208+ cells, there were some activated T-cells that were double positive for CD3 and HLA-DR(Figure 4A–D). In addition, IgA+ cells mainly existed in the T-cell abundant interfollicular area adjacent to the follicles(Figure 4E–H).

Because we found that the increase in CD208+ cells was themost characteristic feature in tonsils of IgAN patients, we in-vestigated the CD208+ cells in renal biopsy tissues. The renalbiopsy specimens of all the 17 IgAN patients whose specimens

F IGURE 1 : Histological structure of tonsils. Original magnification; ×10. (A and B) Representative photomicrograph of HE-stained tonsillartissue. (A) A patient with chronic tonsillitis. (B) A patient with IgA nephropathy (IgAN). Enlarged interfollicular areas and small germinalcenters are characteristic in the tonsils of patients with IgAN. (C) Germinal center area and total tonsil area were measured, and proportions ofgerminal center area to total tonsillar area were calculated. Results are means ± SE for each group. The proportion of germinal center to total ton-sillar area in IgAN patients was significantly smaller than that in controls. *P < 0.01 versus controls. Control: n = 9, IgAN: n = 33.

ORIG

INALARTIC

LE

H. Takechi et al.

4

F IGURE 2 : Various subtypes of DCs in tonsils (A–E) and kidneys (F) of patients with IgAN. Original magnification; A–C, E and F: ×200, D:×100. (A) IF staining for CD303 (under-epithelial area). The arrows show epithelium. (B) IF staining for CD1c (interfollicular area). (C) Immu-noperoxidase staining for CD209 (interfollicular area). The arrow shows a follicle. (D) Immunoperoxidase staining for CD208 (interfolliculararea). The arrow shows a follicle. CD208+ cells are clustered in the interfollicular area. (E) IF staining for CD1a (under-epithelial area). Thearrows show epithelium. (F) IF staining for CD208 in the renal tissue of an IgAN patient. CD208+ cells are observed in the renal interstitium.Photograph shows CD208+ cells around a globally sclerotic glomerulus.

F IGURE 3 : Quantification analysis of different markers of subtypes of DCs in tonsillar tissues from controls and IgAN patients. Results aremeans ± SE for each group. (A) CD303+, (B) CD1c+, (C) CD209+, (D) CD208+ cells. The number of CD208+ cells in the tonsils of IgAN patientswas significantly greater than that in the tonsils of controls. *P < 0.05 versus controls. (E) CD1a+ cells. Control: n = 9, IgAN: n = 33.

ORIG

INALARTIC

LE

I n c r e a s e i n C D 2 0 8 + D C s i n t o n s i l s o f I g A n e p h r o p a t h y

5

were stained for CD208 had CD208+ cells in the interstitium,but in all samples the number of these was very small(Figure 2F). There was no correlation between the number ofCD208+ cells in the kidneys and that in the tonsils.

Flow cytometry analysisTo evaluate DC markers on isolated cells, we examined

fresh tonsillar specimens by flow cytometry. Figure 5A–Fshows representative results of flow cytometry of tonsils froma patient with IgAN, whereas Figure 5G shows a result from acontrol subject. For CD303, CD1c and CD1a, cells that weredouble positive for HLA-DR and DC markers were identifiedto be DCs (Figure 5A–C). For CD209 and CD208, a lineagemarker-negative population (presented as lin− in Figure 5D)was analyzed and is shown in Figure 5E–G. A populationwhich was a lineage marker− and HLA-DR+ and DC marker+

was identified to be DCs.As shown in Figure 5K, the percentage of CD208+ cells in

the lineage marker− and HLA-DR+ population was signifi-cantly greater in the tonsils of IgAN patients than that in thetonsils of controls. The percentage of CD303+, CD1c+,CD209+ or CD1a+ DCs did not differ between the tonsils ofcontrols and those of IgAN patients (Figure 5H–J and L).

Real-time reverse transcription–PCRThe levels of CD208 and CD1c mRNA in the tonsils of

IgAN patients were significantly greater than that in the

tonsils of controls (Figure 6B and D). The mean value ofCD303, CD209 and CD1a mRNA tended to be higher inIgAN patients than in controls, but the differences were notstatistically significant (Figure 6A, C and E).

Positive correlation of DCs with clinical and renalhistological featuresThe numbers of CD208+ cells in the tonsils were signifi-

cantly and positively correlated with the proportion of cres-centic glomeruli in the renal biopsy tissues (R = 0.49, P < 0.05)and with the urinary protein levels (R = 0.38, P < 0.05)(Figure 7A and B). No correlation was found between thenumber of CD208+ cells in tonsils and the level of hematuria,serum creatinine, serum IgA level or the proportions of sclero-tic glomeruli, glomeruli with adhesion or glomeruli with me-sangial proliferation to the total glomeruli. CD303, CD1c,CD209 and CD1a had no correlation with clinical or renal his-tological data.

DISCUSSION

In this study, we found that the number of CD208+ DCs wassignificantly increased in the tonsils of IgAN patients. Thenumber of CD208+ DCs in the tonsils was significantly andpositively correlated with the proportion of crescentic glomer-uli in the renal biopsy tissue and with the urinary protein level.

F IGURE 4 : Triple IF staining of tonsils of patients with IgAN. (A–D) Triple IF staining for CD208 (Alexa Fluor 594: red), CD3 (Alexa Fluor350: blue) and HLA-DR (Alexa Fluor 488: green). (A) CD208, (B) CD3, (C) HLA-DR. HLA-DR is known to be expressed on DCs, macrophages,B-cells and activated T-cells. (D) Merged image of A–C. The arrows indicate activated T-cells (CD3+ HLA-DR+). (E–H) Triple IF staining forCD20 (B-cells; Alexa Fluor 594: red), CD3 (T-cells; Alexa Fluor 350: blue) and IgA (FITC: green). (E) CD20. The arrows indicate follicles. (F)CD3, (G) IgA and (H) merged image of E–G.

ORIG

INALARTIC

LE

H. Takechi et al.

6

Histologically, tonsillar tissues consist of three microcom-partments: crypt epithelium, interfollicular area and lymphoidfollicle [17, 18]. Immune responses against foreign antigensarise in crypt epithelium with the uptake of the antigen by epi-thelial cells. The antigen is then processed and presented byDCs to CD4+ T-cells under the epithelium. The activatedCD4+ T-cells then activate B-cells to induce differentiation ofantigen-specific B-cells in the follicles. B-cells finally differen-tiate into immunoglobulin-secreting plasma cells in the inter-follicular area [19]. Thus, the interfollicular area is animportant place where the antigen-specific T-cell activationand subsequent T-cell–B-cell interaction takes place. Hotta,Takechi et al. [20] and Kawaguchi et al. [21] previously re-ported that the enlarged T-cell nodules in the interfollicularareas of tonsils were characteristic of IgAN patients. In thisstudy, interfollicular areas were characteristically expanded,while areas of germinal centers were reduced in the tonsilsof IgAN.

In the present study, CD1a+ DCs were found under the epi-thelial area, CD303+ DCs were found sparsely in both the epi-thelial area and the interfollicular area. CD1c+, CD209+ andCD208+ DCs were found mainly in the interfollicular area. Wetherefore speculate that CD1c+, CD209+ and CD208+ DCsplay important roles in the expanded interfollicular area in thetonsils of IgAN patients. CD208+ DCs were the most abundantof these DC subtypes and are the characteristic cell type in theinterfollicular area of the tonsils of IgAN patients.

CD208, also known as DC-LAMP, is a member of theLAMP family and known as a specific marker of iDCs [13, 22].It is specifically expressed by human DCs upon activation andmaturation, and CD208+ DCs are generally located in the inter-follicular area of lymphoid tissues. In human DCs, CD208 is ex-pressed in the intracellular major histocompatibility complex(MHC) Class II compartment just before the translocation ofMHC Class II molecules to the cell surface [13]. So, it issuggested that CD208 participates in processing foreign

F IGURE 5 : Flow cytometry analysis of a patient with IgAN (A–F) and a control subject (G). (A–C) Isolated cells were double-stained withantibodies for HLA-DR and either CD303, CD1c or CD1a. The HLA-DR-positive population was analyzed and is shown in graphs. HLA-DR+

and DC-marker+ cells in each staining were identified to be DCs. (A) HLA-DR and CD303 staining. (B) HLA-DR and CD1c staining. (C) HLA-DR and CD1a staining. (D–G) Isolated cells were triple-stained with antibodies for HLA-DR, lineage marker and either CD209 or CD208. Thelineage marker-negative population (indicated lin− in graph D) was analyzed and is shown in E–G. Lineage marker−, HLA-DR+ and DC-marker+ cells in each staining were identified to be DCs. (E) HLA-DR and CD209 staining in the lineage marker− population in a patient withIgAN. (F) HLA-DR and CD208 staining in the lineage marker− population in a patient with IgAN. (G) HLA-DR and CD208 staining in thelineage marker− population in a control subject. (H–L) Quantification analysis results for the different subtypes of DCs in cells isolated fromtonsils of IgAN patients and controls. (H) CD303+ DCs, (I) CD1c+ DCs, (J) CD209+ DCs, (K) CD208+ DCs. The percentage of HLA-DR+ cellsin the lineage marker− population that were CD208+ was significantly greater in the tonsils of IgAN patients than that in the tonsils of controls.*P < 0.05 versus control. (L) CD1a+ DCs. Results are expressed as means ± SE for each group. Control: n = 3, IgAN: n = 4.

ORIG

INALARTIC

LE

I n c r e a s e i n C D 2 0 8 + D C s i n t o n s i l s o f I g A n e p h r o p a t h y

7

antigens in the early stage. In addition to the close relation withMHC Class II molecules, CD208+ DCs strongly express costi-mulatory molecules such as CD80, CD83 and CD40, which areassociated with the adhesion and activation of T-cells [23, 24].Furthermore, the expression of CD208 is correlated with that ofother maturation markers, such as CD86 [24]. We thereforespeculate that CD208+ DCs have a strong ability to activate T-cells. Our observation of an increased number of CD208+ DCsaccompanied by surrounding activated T-cells in an enlargedinterfollicular area suggests that T-cell activation is acceleratedin the tonsils of patients with IgAN. We think that the reductionof relative germinal center areas is a result of the expansion of

interfollicular areas (T-cell areas) that is brought about by thechronically increased number of CD208+ DCs having a strongability to activate T-cells.

Previous reports have shown that, in IgAN, the predomi-nant subset of IgA in mesangial deposits is IgA1, the O-glycanchains in the hinge region of glomerular and serum IgA1 arehighly underglycosylated [25, 26] and the IgA1 molecules pro-duced by tonsillar lymphocytes are also underglycosylated[27]. In the present study, IgA+ cells were found in enlargedinterfollicular areas near follicles. We suspect that, in the de-velopment of IgAN, CD208+ DCs may activate the immunesystem and lead to aberrant IgA1 production in tonsils.

F IGURE 6 : mRNA expression for various markers of DCs. (A) CD303 mRNA, (B) CD1c mRNA, (C) CD209 mRNA, (D) CD208 mRNA and(E) CD1a mRNA. The levels of CD208 and CD1c mRNA in the tonsils of IgAN patients were significantly greater than that in the tonsils of con-trols. Results are expressed as means ± SE for each group. *P < 0.05 versus controls. Control: n = 8, IgAN: n = 23.

F IGURE 7 : Correlation between the number of tonsillar CD208+ DCs and clinical findings. (A) A significant and positive correlation betweenthe number of CD208+ cells in the tonsils and the proportion of crescentic glomeruli to total glomeruli in the renal biopsy tissues of patientswith IgAN (n = 33). (B) A significant and positive correlation between the number of CD208+ cells in the tonsils and the urinary protein level(n = 33). CD303, CD1c, CD209 and CD1a had no correlation with clinical or renal histological data.

ORIG

INALARTIC

LE

H. Takechi et al.

8

Increases in CD208+ DCs in the lesions associated withdermal diseases such as psoriasis [28] and in various types ofcancer, such as colorectal cancer or uterine cervical cancer,have been reported [29, 30]. In the present study, we foundthat the number of tonsillar CD208+ DCs was positively corre-lated with the proportion of crescentic glomeruli in the kidneyand with the level of proteinuria. We, however, found fewCD208+ cells in the renal biopsy tissues of IgAN patients. Wesuspect that the increased CD208+ DCs in the tonsils do notthemselves migrate to the kidney, but may play a directive roleby activating T-cells. Then, the activated T-cells themselvesand/or subsequently produced immunoglobulins from acti-vated B-cells may eventually cause kidney injury in IgAN. Thealternative possibility is that some CD208+ DCs may migratefrom tonsils to the kidney and cause inflammation there, butthey can hardly be identified by the usual histochemical stain-ing of renal biopsy tissues, because there are so few of them inthe kidney.

Several reports suggest that toll-like-receptors (TLRs),especially TLR9, play important roles in the pathogenesis ofIgAN. One study reported a correlation between the tonsillarTLR9 expression level and the therapeutic efficacy of tonsil-lectomy and steroid pulse therapy [31]. Another study re-ported that the expression of TLR9 and that of its ligand,MyD88, in spleen were correlated with the glomerular injuryin animal models of IgAN [8]. Because the CD303+ DC(pDC) is the main DC that expresses TLR9 [32], it had beenexpected that CD303+ DCs would be increased in the tonsilsof patients with IgAN. In the present study, however, we didnot find such an increase. Since TLR9 is also on B- and T-cellsbesides CD303+ DCs in human [33, 34], in IgAN it may beoverexpressed on these cells rather than CD303+ DCs.

The numbers of myeloid-related DCs in nasal-associatedlymphoid tissue were previously reported to be increased inpatients with IgAN: the numbers of CD1a+ DCs and CD209+

DCs were increased in the nasal biopsy samples of patientswith IgAN [35]. We, however, did not find increases of CD1a+

or CD209+ DCs in the tonsils of IgAN patients. This isprobably due to the difference between the DC populationsof tonsillar tissues, which are lymphoid tissues, and nasalmucosa.

In summary, CD208+ DCs were significantly increased inthe tonsils of patients with IgAN, and the number of tonsillarCD208+ DCs was positively correlated with the proportion ofglomerular crescent formation and the level of proteinuria.The present results suggest the possibility that increasedmyeloid-related mature DCs, CD208+ DCs, reinforce theinnate and adaptive immunity, thereby resulting in deterio-ration of renal lesions. The present results, thus, may supportthe therapeutic significance of tonsillectomy for IgANpatients. Further studies are required to find out how tonsillarCD208+ DCs induce renal lesion in patients with IgAN.

ACKNOWLEDGEMENTS

We are grateful to our colleagues, Ms Tatsuyo Harasawa (atthe Central Research Laboratory of National Defense Medical

College) and Ms Sachiko Iwama (a technologist in our Depart-ment) for excellent technical assistance and to Dr Shuhji Seki,Dr Hiroyuki Nakashima (Department of Immunology andMicrobiology, National Defense Medical College), Dr KensukeJoh (Department of Pathology, Sendai Shakaihoken Hospital)and Dr Hiroki Sato (Department of Preventive Medicine andPublic Health, National Defense Medical College) for valuableadvice and discussion.

CONFLICT OF INTEREST STATEMENT

None declared.

REFERENCES

1. Emancipator SN, Lamm ME. IgA nephropathy: pathogenesis ofthe most common form of glomerulonephritis. Lab Invest 1989;60: 168–183

2. Levy M, Berger J. Worldwide perspective of IgA nephropathy.Am J Kidney Dis 1988; 12: 340–347

3. Manno C, Strippoli GF, D’Altri C et al. A novel simpler histologi-cal classification for renal survival in IgA nephropathy: a retro-spective study. Am J Kidney Dis 2007; 49: 763–775

4. Hotta O, Miyazaki M, Furuta T et al. Tonsillectomy and steroidpulse therapy significantly impact on clinical remission inpatients with IgA nephropathy. Am J Kidney Dis 2001; 38:736–743

5. Komatsu H, Fujimoto S, Hara S et al. Effect of tonsillectomy plussteroid pulse therapy on clinical remission of IgA nephropathy: acontrolled study. Clin J Am Soc Nephrol 2008; 3: 1301–1307

6. Kennoki T, Ishida H, Yamaguchi Y et al. Proteinuria-reducingeffects of tonsillectomy alone in IgA nephropathy recurring afterkidney transplantation. Transplantation 2009; 88: 935–941

7. Ushigome H, Suzuki T, Fujiki M et al. Efficacy of tonsillectomyfor patients with recurrence of IgA nephropathy after kidneytransplantation. Clin Transplant 2009; 23: 17–22

8. Kajiyama T, Suzuki Y, Kihara M et al. Different pathological rolesof toll-like receptor 9 on mucosal B cells and dendritic cells inmurine IgA nephropathy. Clin Dev Immunol 2011; 2011: 819646

9. Goto T, Bandoh N, Yoshizaki T et al. Increase in B-cell-activationfactor (BAFF) and IFN-gamma productions by tonsillar mono-nuclear cells stimulated with deoxycytidyl-deoxyguanosine oligo-deoxynucleotides (CpG-ODN) in patients with IgA nephropathy.Clin Immunol 2008; 126: 260–269

10. Gilliet M, Cao W, Liu YJ. Plasmacytoid dendritic cells: sensingnucleic acids in viral infection and autoimmune diseases. Nat RevImmunol 2008; 8: 594–606

11. Colonna M. Skin function for human CD1a-reactive T cells. NatImmunol 2010; 11: 1079–1080

12. Noessner E, Lindenmeyer M, Nelson PJ et al.. Dendritic cells inhuman renal inflammation—part II. Nephron Exp Nephrol 2011;119: e91–e98

13. de Saint-Vis B, Vincent J, Vandenabeele S et al. A novel lyso-some-associated membrane glycoprotein, DC-LAMP, inducedupon DC maturation, is transiently expressed in MHC class IIcompartment. Immunity 1998; 9: 325–336

ORIG

INALARTIC

LE

I n c r e a s e i n C D 2 0 8 + D C s i n t o n s i l s o f I g A n e p h r o p a t h y

9

14. Hyodo T, Oda T, Yamamoto K et al. Voltage-gated potassiumchannel Kv1.3 blocker as a potential treatment for rat anti-glomerular basement membrane glomerulonephritis. Am JPhysiol Renal Physiol 2010; 299: F1258–F1269

15. Angel CE, Chen CJ, Horlacher OC et al. Distinctive localizationof antigen-presenting cells in human lymph nodes. Blood 2009;113: 1257–1267

16. Yamada M, Oda T, Higashi K et al. Involvement of epimorphinin the repair of experimental renal fibrosis in mice. Lab Invest2010; 90: 867–880

17. Nave H, Gebert A, Pabst R. Morphology and immunology of thehuman palatine tonsil. Anat Embryol (Berl) 2001; 204: 367–373

18. Xie Y, Chen X, Nishi S et al. Relationship between tonsils andIgA nephropathy as well as indications of tonsillectomy. KidneyInt 2004; 65: 1135–1144

19. Hoefakker S, van ’t Erve EH, Deen C et al. Immunohistochemicaldetection of co-localizing cytokine and antibody producing cellsin the extrafollicular area of human palatine tonsils. Clin ExpImmunol 1993; 93: 223–228

20. Takechi H, Nagura H, Houzawa K et al. Clinical and immuno-logical study of the tonsils in IgA nephropathy. Contrib Nephrol2007; 157: 197

21. Kawaguchi M, Sakai T, Sakamaki A et al. Expanded primary Tnodules in the palatine tonsils from patients with IgA nephropa-thy. Acta Otolaryngol Suppl 1993; 508: 36–42

22. Steinman RM, Hemmi H. Dendritic cells: translating innate toadaptive immunity. Curr Top Microbiol Immunol 2006; 311:17–58

23. Zhou LJ, Schwarting R, Smith HM et al. A novel cell-surface mol-ecule expressed by human interdigitating reticulum cells, Langer-hans cells, and activated lymphocytes is a new member of the Igsuperfamily. J Immunol 1992; 149: 735–742

24. Bjorck P, Flores-Romo L, Liu YJ. Human interdigitating dendriticcells directly stimulate CD40-activated naive B cells. Eur JImmunol 1997; 27: 1266–1274

25. Hiki Y, Odani H, Takahashi M et al. Mass spectrometry provesunder-O-glycosylation of glomerular IgA1 in IgA nephropathy.Kidney Int 2001; 59: 1077–1085

26. Odani H, Hiki Y, Takahashi M et al. Direct evidence for de-creased sialylation and galactosylation of human serum IgA1 FcO-glycosylated hinge peptides in IgA nephropathy by mass spec-trometry. Biochem Biophys Res Commun 2000; 271: 268–274

27. Horie A, Hiki Y, Odani H et al. IgA1 molecules produced by ton-sillar lymphocytes are under-O-glycosylated in IgA nephropathy.Am J Kidney Dis 2003; 42: 486–496

28. Higaki M, Higaki Y, Kawashima M. Increased expression ofCD208 (DC-LAMP) in epidermal keratinocytes of psoriaticlesions. J Dermatol 2009; 36: 144–149

29. Kanao H, Enomoto T, Kimura T et al. Overexpression ofLAMP3/TSC403/DC-LAMP promotes metastasis in uterine cer-vical cancer. Cancer Res 2005; 65: 8640–8645

30. Sandel MH, Dadabayev AR, Menon AG et al. Prognostic value oftumor-infiltrating dendritic cells in colorectal cancer: role ofmaturation status and intratumoral localization. Clin Cancer Res2005; 11: 2576–2582

31. Sato D, Suzuki Y, Kano T et al. Tonsillar TLR9 expression and ef-ficacy of tonsillectomy with steroid pulse therapy in IgA nephro-pathy patients. Nephrol Dial Transplant 2012; 27: 1090–1097

32. Akira S, Uematsu S, Takeuchi O. Pathogen recognition andinnate immunity. Cell 2006; 124: 783–801

33. Mansson A, Adner M, Cardell LO. Toll-like receptors in cellularsubsets of human tonsil T cells: altered expression during recur-rent tonsillitis. Respir Res 2006; 7: 36

34. Krieg AM, Vollmer J. Toll-like receptors 7, 8, and 9: linking innateimmunity to autoimmunity. Immunol Rev 2007; 220: 251–269

35. Eijgenraam JW, Reinartz SM, Kamerling SW et al. Immuno-histological analysis of dendritic cells in nasal biopsies of IgAnephropathy patients. Nephrol Dial Transplant 2008; 23: 612–620

Received for publication: 17.4.2013; Accepted in revised form: 9.8.2013

ORIG

INALARTIC

LE

H. Takechi et al.

10